INDEX

Explanations

proper nouns related to names of people or places

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the token "ne"

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rador

-1.01

hips

-0.87

rament

-0.85

inarily

-0.83

rican

-0.80

enhagen

-0.78

allery

-0.77

IAL

-0.76

orsi

-0.75

redited

-0.74

POSITIVE LOGITS

arest

1.02

zel

0.90

cker

0.86

theless

0.86

gan

0.84

jad

0.84

cks

0.83

phrine

0.83

cht

0.83

verend

0.83

Activations Density 0.020%

proper nouns related to names of people or places

repeated instances of the token "ne"

No Comments

No Known Activations